Problem w new Mitutoyo 20x. Looking for community feedback.

[abpho]

Hello and thanks for looking.

I purchased a new Mitutoyo M plan 20x 0.42 lens being sold as surplus stock. The lens is in immaculate condition, verifying it's stated condition. I do see some dust inside the lens when using a flashlight. I was not too worried about this. Dust happens. This is the latest version of the lens. The part number is 378-804-3.

I have been taking images for the past day and a half. I have provided links to the images on my flickr account.

NOTE: Having issues posting topic. Trying to break into smaller parts.

[abpho]

First I shot a stack of a butterfly wing using a Nikon BD Plan 40x 0.65 210/0 on a full frame camera. 0.001mm steps. 172 images. Flash illuminated. Horizontal field of view is 0.9mm. The edited in post full resolution image is here.

I then photographed the same area with the Mitutoyo M Plan 20x 0.42 ?/0 f=200 on a Canon EF 200mm f/2.8L USM II (focused to infinity) using a Canon EF 1.4x II Teleconverter on a APS-C sensor size camera. 0.001mm steps. 141 images. Flash illuminated. Horizontal field of view is 0.8mm. The stacked full resolution image is here.

With the teleconverter and the smaller sensor size I am shooting with the same field of view. I was expecting a huge improvement in resolving power compared the Mitutoyo. I have the Mitutoyo 5x and 10x lenses, and I have seen what they are capable of.

NOTE: I think the editor did not like me using the infinity symbol in the Mitutoyo's description. It was substituted for a ? above. Sorry for the multiple posts.

[abpho]

Finally I shot a stack of the same area with a Nikon M Plan 20x 0.4 210/0 ELWD on a full frame camera. 0.002mm steps. 127 images. Flash illuminated. Horizontal field of view is 1.8mm. The edit image was stretched by a factor of 2. That image can be found here.

So far the stack with the 40x shows the most amount of detail. Just to make sure sure I used the Mitutoyo 20x again on it's own with a full frame camera. Same results.

I was thinking perhaps the subject was not the greatest. I just happened to have one bug left in the freezer from a few years back. I took a small 61 image stack. 0.002mm increments. Flash illuminated. Here is a 100% crop taken from the centre of the image. Not good in my books.

I would like to hear your comments before contacting the seller and asking for a refund or an exchange.

Thank you.


Mischa.

[rjlittlefield]

I purchased a new Mitutoyo M plan 20x 0.42 lens being sold as surplus stock.
...
First I shot a stack of a butterfly wing using a Nikon BD Plan 40x 0.65 210/0 on a full frame camera. 0.001mm steps. 172 images. Flash illuminated. Horizontal field of view is 0.9mm. The edited in post full resolution image is here.

I then photographed the same area with the Mitutoyo M Plan 20x 0.42 ?/0 f=200 on a Canon EF 200mm f/2.8L USM II (focused to infinity) using a Canon EF 1.4x II Teleconverter on a APS-C sensor size camera. 0.001mm steps. 141 images. Flash illuminated. Horizontal field of view is 0.8mm. The stacked full resolution image is here.

With the teleconverter and the smaller sensor size I am shooting with the same field of view. I was expecting a huge improvement in resolving power compared the Mitutoyo.
...
I would like to hear your comments before contacting the seller and asking for a refund or an exchange.

OK, I'm confused.

It sounds like you're comparing an NA 0.65 objective with an NA 0.42 objective, and expecting the NA 0.42 to have more resolution on subject because you've stuck a teleconverter and a small sensor behind it.

That sounds misguided, so surely I must have misunderstood.

Can you please tell us again, in different words, what you're expecting and why?

--Rik

[abpho]

Hi Rik.

I'll try to explain myself.

First and foremost, I don't think about the magnification of a lens. I think about the field of view I need to capture my subject. I don't have any prior microscope experience. I am riding on the coattails of others. I see what they are capable of capturing and I base my knowledge on that.

I took the first image with the 40x objective as my base line. As I have stated, that particular setup netted a horizontal field of view of 0.9mm. I have seen others capture stunning results using various Mitutoyo lenses above and below their stated magnifications. (I currently do not have any specific examples to show you.) That is why I took my next picture with the Mits 20x on a crop sensor and a 1.4x teleconverter. The field of view was close to the first image at 0.8mm. That's all I care about.

Seeing that the results were much worse than what I expected (from my own experience with Mitutoyo lenses, and from other people's works as well) I took two more pictures at the stated 20x magnification with the Mitutoyo 20x and the Nikon M Plan 20x 0.4 210/0. Even in this case I felt the Mitutoyo is not living up to it's potential.

Then I changed the subject. And again the Mitutoyo seems to be under performing.

All I can base my expectation on is what I have seen with the Mits 5x and 10x compared to my other lenses. I was always blown away with the detail those two could capture. Am I expecting too much?

I hope this makes more sense. Sorry for my ignorance.

[Beatsy]

Looking at your butterfly scales pic, that 20x seems to be performing very well considering.

With the 200mm tube lens and 1.4x extender, magnification onto the sensor is 28x. The smallest distance the 20x can resolve (given the N.A.) is 0.7 microns. So that distance covers 28*0.7 = 19.6 microns on sensor. So if your pixels are 5 microns then the smallest resolveable area covers 4 pixels. That's going to look soft in a 1:1 view on screen.

[abpho]

Thanks Beatsy. But what about bigger objects. Here is the last image, but in full resolution and sharpened.

I just want to make sure this lens is working properly. I have no problem with the expense. I do have a problem if the Mits is not up to it's full potential and I end up with a $1.55k paper weight.

[Beatsy]

Still looks OK to me (assuming you're on 200mm with 1.4x extender).

That's a very difficult subject though, hard to get a good idea of resolution and lens basic performance from such a reflective critter.

I'd dump the 1.4x extender - it's only adding empty magnification IMO. You get the same effect by cropping into a 20x image in post (just using 200mm tube lens for capture). For online pics, I shrink the image by 50% in Zerene while stacking too. This sharpens things up no end and takes a lot less time to stack.

Oh - and clean that sensor

Cheers
Beats

[abpho]

I only used the 1.4x in that one butterfly image. For this bug picture it's just the Mits 20x, a Canon EF 200L 2.8, and a Canon EOS 5D2.

[Beatsy]

I still think it's OK.

A surer quick test is that butterfly wing. Take one stack with your 10x mitty and then another, of the same area under the same lighting, with the 20x. Now, crop the 10x image to cover the same area as the 20x and then shrink the 20x image by 50%. You should then have two near-identical images. Blink between the two to compare them. The 20x should still look better, but only a bit (0.7 micron resolution versus the 1 micron resolution of the 10x).

[abpho]

Thanks Beatsy. I will do that tonight.

And a quick history lesson....Lol...I had to go through three Mitutoyo 10x lenses. It was too much to deal with. I don't want a repeat of that. Sometimes buying used doesn't pay. Luckily my Mitty 5x had no issues.

[rjlittlefield]

All I can base my expectation on is what I have seen with the Mits 5x and 10x compared to my other lenses. I was always blown away with the detail those two could capture. Am I expecting too much?

Perhaps a bit of theory will help...

The maximum resolution of any lens, measured on subject, is just proportional to its NA. This limit is imposed by diffraction. So, comparing NA 0.42 and NA 0.65 lenses, the NA 0.65 has potentially better resolution in the ratio of 0.65/0.42 = 1.55 .

Lenses cannot do any better than the diffraction limit. They can do arbitrarily worse due to various aberrations. However, high quality microscope objectives in good condition will perform very close to the theoretical limits at the center of their fields, except for longitudinal CA.

So, there is every reason to expect that a Nikon BD Plan NA 0.65 should be significantly sharper on subject than a Mitutoyo M Plan Apo NA 0.42, assuming that both lenses are in good condition. If that turns out not to be true in practice, then either the test was flawed or the NA 0.65 objective is a duff sample.

Similarly, comparing NA 0.42 and NA 0.28, the NA 0.42 has potentially better resolution in the ratio of 0.42/0.28 = 1.50 . And so, your Mitutoyo M Plan Apo NA 0.42 should be significantly sharper on subject than your Mitutoyo M Plan Apo NA 0.28 .

In both cases, it may be tricky to do a valid comparison because the objectives are designed to give different magnifications. Usually the best approach is to use the finest pitch sensor that you have, and push up the effective magnification of the lower power objective, optically if possible (for example using a teleconverter) or digitally if necessary (by resampling to more pixels), or a combination of those two, so that in the end you cover the same size field with the same number of pixels. Then you can do a direct visual comparison, similar to images that you have posted.

Note that pushing the magnification up is what you do to get a valid comparison between two lenses. That works because it makes less likely that sensor resolution will degrade the image. The downside is that this makes the test image blurred because every feature gets smeared across several pixels. To make the sharpest real photographs, you go the other direction and push the magnification down so that each feature gets smeared over as few pixels as possible.

To get the sharpest picture of any particular subject, what you want is the largest NA objective that has a field sufficient to cover the subject, and you want the image to cover the sensor as closely as possible.

So, the perfect subject for your 20X NA 0.42 objective is one that is too big to cover with your 40X NA 0.65, but small enough that you can cover it with the NA 0.42, pushed down if necessary by using a shorter tube lens.

As an example, the sharpest 10X in my arsenal, for an APS-C sensor, is actually a Mitutoyo 20X NA 0.42 pushed down to 10X with a 100 mm tube lens. The corners are not as good as the center, but they're still quite good compared to the 10X at smaller NA.

I do have a problem if the Mits is not up to it's full potential and I end up with a $1.55k paper weight.

Yep, you don't want that. I confess, given your history with used objectives, I'm surprised that you went for a used 20X at that price. List price new from Edmund Optics is only $2.05k.

I hope this helps. Run those tests suggested by Beatsy.

--Rik

[abpho]

Thank you Rik for the explanation. I will try to digest that.

I confess, given your history with used objectives, I'm surprised that you went for a used 20X at that price. List price new from Edmund Optics is only $2.05k.

That is $2.05k US. I am talking Canadian. If you convert $2.05KUS to Canadian you get roughly $2690. Then don't forget to throw in a little shipping and handling too.

Canada sucks.

Currently working on Beatsy suggestion.

[abpho]

Okay. I ran the test suggested by Beatsy.

Here is the link to the cropped image taken with the Mitutoyo M Plan 10x 0.28.

Here is the link to the resized image taken with the Mitutoyo M Plan 20x 0.42.

Here is the link to the animated GIF.

What are your thought? There is a small difference. I am almost leaning towards the Mitty 20x resolving ever so slightly more detail.

[Beatsy]

Definitely more resolution showing in the 20x IMO. Look where the "ladders" are clear in the white scales. They're all definitely slightly fuzzier in the 10x image.

Looks about right to me. Congratulations on your purchase